Multi-Band Sensing in FR3 with Background Dense Multipath Components

arXiv:2602.23539 · eess.SP · Submitted 2026-02-26 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Electrical Engineering. The day in electrical engineering: signals, systems and the hardware that carries them.

Rosa: I'm Rosa.

Dev: And I'm Dev, here with Taro.

Rosa: We're looking at the paper: "Multi-Band Sensing in FR3 with Background Dense Multipath Components".

Dev: Here's the gist of the paper. Multi-band sensing has emerged as a key enabler of integrated sensing and communication (ISAC), one of the six primary usage scenarios defined for IMT-2030 (6G) The paper proposes a novel ISAC channel analysis tailored to multi-band sensing, based on a channel model with background DMCs It assesses the sensing trade-offs among sub-bands by analyzing Cramer-Rao bound (CRB)-based fundamental limits Furthermore, a scalable multiband estimator is proposed that resolves angular ambiguities arising from the grating lobes effect.

Rosa: Multi-band sensing is the technique of performing sensing across non-contiguous frequency bands The authors discuss reasons for choosing multi-band sensing, such as achieving higher time resolution and mitigating loss from severe fading at a particular frequency band.

Dev: A key propagation effect is the dense multipath components (DMC), which refers to clutter, such as diffuse scattering and diffraction, from the background environment.

Rosa: DMC is frequency-dependent and should be modeled differently across sub-bands In multi-band settings, DMC becomes even more determinant because its statistics are frequency-dependent.

Title and authors: Rosa: The authors, Dexin Wang, Roberto Bomfin, Ahmad Bazzi, and Marwa Chafii are taking this multi-band sensing concept into the real world for 6G <ref:2602.23539#pg1,Dexin Wang, Roberto Bomfin, Ahmad Bazzi, and Marwa Chafii>.

Dev: They’re proposing a new way to analyze the channel specifically for this multi-band setup by building their model around these background dense multipath components.

Taro: And they aren't just saying it works; they are looking at the fundamental limits of what we can achieve by analyzing the Cramer-Rao bound, or CRB, for each sub-band individually to see how those trade-offs play out.

Rosa: It’s about showing that you can’t just look at one band and think it's fine; you have to consider how all those different frequency bands interact when you're trying to get a good sense of where things are.

Dev: And they propose this scalable estimator, which is meant to clean up the confusion caused by grating lobes, which is that angular ambiguity problem that pops up in multi-band systems.

Taro: So what this means practically is that if you're using something like FR3 for sensing, you can get better localization accuracy because the new method handles those tricky angle ambiguities more gracefully than older methods.

The paper's summary: Dev: The core message here is that just looking at the reflected signal power isn’t reliable when you have dense multipath components present.

Rosa: They point out that this interference grows as the signal gets stronger, and it can actually hide the real information you're trying to estimate, which leads to really bad results if you ignore it.

Taro: And they emphasize that in a multi-band setup, these multipath statistics are frequency-dependent, making the problem even more complicated than if you only looked at one band.

Rosa: They built a channel model where the channel is broken down into things like scatterers and this background dense multipath component, which they call DMC.

Dev: They then look at how you can analyze the sensing trade-offs among different sub-bands by using something called the BGG metric, which basically separates the geometry interaction from noise and interference.

Taro: It’s a way to get a more physically meaningful view of performance that isn't just about raw signal strength.

Rosa: On top of that, they introduce this multi-band estimator that combines estimates from single bands using weights based on the CRB to resolve those angular ambiguities caused by the grating lobes effect.

Dev: So, in short, it’s a new channel analysis method combined with a new estimator designed to give you better accuracy and lower false alarm rates compared to just running single-band estimators separately.

The paper's improvements: Taro: The main improvement they bring is this scalable multi-band estimator that solves the angular ambiguity problem that comes from grating lobes.

Rosa: They show simulation results where this combined approach beats single-band estimators significantly in terms of accuracy and false alarm rate reduction.

Dev: Specifically, in one test case, their estimator cuts the root mean squared error of delay estimation by thirty-seven point four one percent when compared to a single-band estimator operating at eight point seven five gigahertz.

Rosa: And that’s not all; they show performance curves where the proposed multi-band estimator has a much lower false alarm rate than any of the single-band estimators on their constituent sub-bands.

Dev: This is really interesting because it suggests that combining the bands in this specific way leads to better overall reliability when you're trying to localize something.

Taro: The method itself uses a linear combination of estimates from different sub-bands, and they use a Hungarian matching algorithm to make sure they correctly link the estimates across all those bands and eliminate the aliased angles.

Conclusion: Rosa: To wrap up, this paper shows that accurately modeling the background dense multipath components is absolutely critical for multi-band ISAC in future 6G applications <ref:2602.23539#pg1>.

Dev: They confirm that their proposed estimator achieves performance near the fundamental limits when you are operating in the CRB regime and provides real accuracy gains over running single-band estimators on each sub-band.

Taro: The work suggests that this framework is a solid step toward making these multi-band systems practical for things like FR3.

Rosa: It’s about understanding how to get meaningful sensing conclusions when you have all these frequency effects happening at once, and they did that by isolating the geometry interaction with the Bistatic Geometric Gain metric.

Dev: They also plan to develop a joint multi-band algorithm that can achieve the exact Cramer-Rao bound, which is a big goal for getting perfect performance.

Taro: For me, I think the real implication here is that we need to start thinking about how this estimation framework actually runs in real world measurement campaigns so we can verify what they claim.

Rosa: Exactly. It’s not just a theoretical result; it’s about seeing if this works when the world misbehaves and things get messy.

Dev: And from an engineering standpoint, we need to see how fast this estimator can run in real time because latency matters for any control loop you might be building.

Taro: Right, so we know the math looks promising for localization, but now it’s about testing if it survives the actual noise and complexity of the environment.

New York University Abu Dhabi · NYU WIRELESS

eess.SP

Submitted: 2026-02-26

Updated: 2026-11-13

Comments: Accepted for publication in IEEE Transactions on Wireless Communications, 2026

DOI: 10.1109/TWC.2026.3754549

License: http://creativecommons.org/licenses/by-nc-nd/4.0/

Importance score: 98/100

The gist: Multi-band sensing has emerged as a key enabler of integrated sensing and communication (ISAC), one of the six primary usage scenarios defined for IMT-2030 (6G) The paper proposes a novel ISAC…

Key concepts

Background DMC
These are clutter components like diffuse scattering and diffraction in the environment that are present across multiple frequency bands. They are modeled as frequency-dependent interference that grows with transmit power, requiring special handling in multi-band sensing.
Cramer-Rao Bound (CRB)
The CRB provides a fundamental lower limit on the estimation error variance for any parameter, like angle or delay. Analyzing CRBs helps determine the theoretical performance limits and trade-offs when using different frequency sub-bands.
Grating Lobes Effect
This is an angular ambiguity that occurs in multi-band systems, where signals from different frequency bands cause aliasing. The proposed algorithm uses a matching process to resolve these overlapping or confused angle estimates.
BGG Metric
This metric combines the Signal Component (SC), DMC, and noise to isolate the geometry-dependent interaction between these elements. It serves as a physically meaningful measure to guide resource allocation across different frequency bands.

Terminology

Summary

Multi-band sensing has emerged as a key enabler of integrated sensing and communication (ISAC), one of the six primary usage scenarios defined for IMT-2030 (6G) The paper proposes a novel ISAC channel analysis tailored to multi-band sensing, based on a channel model with background DMCs It assesses the sensing trade-offs among sub-bands by analyzing Cramer-Rao bound (CRB)-based fundamental limits Furthermore, a scalable multiband estimator is proposed that resolves angular ambiguities arising from the grating lobes effect. p. 0

Where the paper says it

Building on prior literature and our experimental observations, this paper proposes a novel ISAC channel analysis tailored to multi-band sensing, based on a channel model with background DMCs. It also assesses the sensing trade-offs among sub-bands by analyzing Cramer-Rao bound ´ (CRB)-based fundamental limits.

Background on Multi-band Sensing and DMC

Multi-band sensing is the technique of performing sensing across non-contiguous frequency bands The authors discuss reasons for choosing multi-band sensing, such as achieving higher time resolution and mitigating loss from severe fading at a particular frequency band. A key propagation effect is the dense multipath components (DMC), which refers to clutter, such as diffuse scattering and diffraction, from the background environment. p. 0

Where the paper says it

However, an important propagation effect widely studied in channel modeling (CM) literature, while remaining largely underexplored in ISAC, is the dense multipath components (DMC) [26]–[31]. DMC refers to clutter, such as diffuse scattering and diffraction, from the background environment (and also scatterers) [26].

DMC is frequency-dependent and should be modeled differently across sub-bands In multi-band settings, DMC becomes even more determinant because its statistics are frequency-dependent. p. 1
Where the paper says it

From a signal processing perspective, it introduces non-negligible estimation bias and unexpectedly large errors, and yields erroneous internal reliability metrics for the estimated parameters. In multi-band settings, DMC becomes even more determinant because its statistics are frequency-dependent.

System Modeling and Channel Components

The system model involves a dual-function radar communication (DFRC) base station serving as the transmitter (Tx) and a passive radio (PR) as the receiver (Rx) The multi-band MIMO channel transfer function is expressed as h = s + d + w, where s represents the SC associated with the K paths, d ∼ CN (0,R) represents the background DMC, and w ∼ CN (0,S) represents the noise. The SC steering vector is given by sm = X Σ k=1 gk,mam(τk, ϕk, θk), where am(τk, ϕk, θk) is the total steering vector for path k at sub-band m. p. 2

Where the paper says it

, not changing across different realizations of the channel. Hence, each sm ∈ C NLTL R×1 represents the specular channel TF for each sub-band with frequency fm, which is given by sm = X K k=1 gk,mam(τk, ϕk, θk), (2) where am(τk, ϕk, θk) is the total steering vector for path k at sub-band m described by the geometric parameters, and gk,m is the path gain for path k at sub-band m.

The path gains are modeled using the two-way bistatic radar equation for scatterers and the Friis’ equation for the LoS path. p. 2
Where the paper says it

Thus, we opted for the following model, which absorbs many parameters, including λ 2 m/4π, the RCS of the scatterers σ RCS k,m(∀k > 1), the Tx and Rx antenna directivity patterns u T m(ϕ) and u R m(θ), and phase terms such as e −jωmτk into the complex path coefficient γk,m. The gain of the LoS path (k = 1) is modeled by the Friis’ equation g1,m = γ1,ms PTf∆,m 4π(cτ1) 2 , (7) where |γ1,m| = r λ2 mu T m(ϕ)u R m(θ) 4π , (8) and the complex path gain for each scatterer’s SC gk,m (k > 1) is modeled by the two-way bistatic radar equation gk,m = γk,ms PTf∆,m (4π) 2(cτ D k ) 2(cτ A k ) 2 , (9) where |γk,m| = s λ2 mσ RCS k,mu T m(ϕ)u R m(θ) 4π .

Fundamental Limits and Trade-offs

The paper analyzes sensing trade-offs by analyzing CRB-based fundamental limits The approximate CRB takes the form CRBapprox ψk = X Σ M m=1 CRB−1 ψk,m!−1. The BGG metric is defined as Γm(p) = GSC m(p) GDMC m (p) + σ2 m. p. 5

Where the paper says it

This consequence motivates the introduction of a metric that explicitly accounts for the combined contributions of SC, DMC, and noise on sensing performance. To this end, we define the BGG metric for the mth sub-band as Γm(p) = GSC m(p) GDMC m (p) + σ 2 m , (35) where p denotes the Cartesian coordinates of a 2D location.

This metric isolates the geometry-dependent interaction between SC, DMC, and noise by factoring out path-dependent coefficients γk,m>

Proposed Multi-band Estimation Algorithm

A scalable multi-band estimator is proposed that resolves angular ambiguity caused by the grating lobes effect. The algorithm performs a linear combination of the estimated geometric parameter values from single-band estimators using weights wψˆk,m = CRB−1 ψˆk,m P ∀m detected CRB−1 ψˆk,m. p. 6

Where the paper says it

When the estimators for every sub-band are efficient (i. , they approach their own CRB), we can obtain the minimal variance linear unbiased estimator by designing weights (for any sub-band m where some geometric parameter ψk is estimated) as [41] wψˆk,m = CRB−1 ψˆk,m P ∀m detected CRB−1 ψˆk,m (42) (see Appendix A for detailed derivations).

The Hungarian matching algorithm is used to simultaneously associate estimates across sub-bands and eliminate the aliased estimated angles. p. 6
Where the paper says it

This step is added after we observed that in low SNR regions, where the CRB is not achieved, combining harms estimation accuracy, as it may make the multi-band estimator even less accurate than the better single-band estimator. Matching by Hungarian Algorithm To simultaneously associate estimates across sub-bands and eliminate the aliased estimated angles, we use the Hungarian matching algorithm.

Simulation Results and Performance Gains

Simulation results demonstrate substantial gains in estimation accuracy and reductions in false alarm rate over single-band estimators operating on each constituent sub-band within the CRB-achieving regime. In a representative test case, the proposed estimator achieves reductions of 37. p. 0

Where the paper says it

Simulation results of the multi-band estimator demonstrate substantial gains in estimation accuracy and reductions in false alarm rate over single-band estimators operating on each constituent sub-band within the CRB-achieving regime. In a representative test case, the proposed estimator achieves reductions of 37.

41% and 17.04% in the root mean squared error of delay estimation compared to single-band estimators operating at 8. p. 0
Where the paper says it

41% and 17. 04% in the root mean squared error of delay estimation compared to single-band estimators operating at 8.

75 GHz and 21. p. 0
Where the paper says it

1 Multi-Band Sensing in FR3 with Background Dense Multipath Components Dexin Wang, Roberto Bomfin, Ahmad Bazzi, and Marwa Chafii Abstract—Multi-band sensing has emerged as a key enabler of integrated sensing and communication (ISAC), one of the six primary usage scenarios defined for IMT-2030 (6G). The introduction of frequency range 3 (FR3, 7-24 GHz), comprising non-contiguous sub-bands across a wide frequency span, further reinforces the importance of multi-band operation.

7 GHz, respectively. p. 0
Where the paper says it

75 GHz and 21. 7 GHz, respectively.

The PD-PFA performance curves show that the combined multi-band estimator has a significantly lower PFA compared to any of the single-band estimators on the constituent sub-bands. p. 1
Where the paper says it

Then, to know how to perform multi-band sensing under DMC: • We propose a multi-band estimator capable of resolving angular ambiguity caused by the grating lobes effect. • We observe fundamental limits-achieving performances, with enhanced performance compared to single-band systems in terms of lower root mean squared error (RMSE) and probability of false alarm (PFA).

Conclusion and Future Work

This work establishes the critical role of background DMC modeling in multi-band ISAC for future 6G applications The paper concludes that the proposed estimator achieves near-fundamental limit performance in the CRB-achieving regime and provides substantial accuracy gains and reductions in the false alarm rate over single-band estimators. Future work plans include using the proposed framework in upcoming measurement campaigns to verify the multiband estimator’s capabilities practically. The authors also plan to develop a joint multi-band algorithm that achieves the exact CRB. p. 4

Where the paper says it

The Approximate CRB In Section IV, we propose a multi-band estimator based on CRB-weighted combining. To assess the performance of this estimator, we need to derive its CRB, which is an approximation to the exact joint CRB derived in the previous subsection.

Appendix A Derivations

For unbiased estimators, the optimal weights are derived as wm = σ−2m PM m=1 σ−2m. When individual estimators are efficient, the CRB is equivalently the lower bound of the variances. p. 6

Where the paper says it

When the estimators for every sub-band are efficient (i. , they approach their own CRB), we can obtain the minimal variance linear unbiased estimator by designing weights (for any sub-band m where some geometric parameter ψk is estimated) as [41] wψˆk,m = CRB−1 ψˆk,m P ∀m detected CRB−1 ψˆk,m (42) (see Appendix A for detailed derivations).

This leads to the approximate CRB as in (41) which serves as a lower bound to the exact singleband CRBs for any sub-band. p. 4
Where the paper says it

Also, since all CRB terms are positive, we can notice that X M m′=1 CRB−1 ψk,m′ ≥ CRB−1 ψk,m, ∀m. (32) Inverting both sides and using the definition in (31), we obtain CRBapprox ψk ≤ CRBψk,m, ∀m, (33) suggesting that the approximate multi-band CRB discussed in this subsection serves as a lower bound to the exact singleband CRBs for any sub-band.

The objective function for minimizing MSE is equivalent to minimizing the variance of ψˆ, leading to wm ∝ 1/σ2m.

References

[1] Z. Tao, W. Xu, and X. You, “Provable Performance Bounds for Digital Twin-Driven Reinforcement Learning in Wireless Networks: A Novel Digital-Twin Bisimulation Metric,” IEEE Transactions on Signal Processing, vol. p. 0

Where the paper says it

INTRODUCTION F uture 6G systems are envisioned to reliably facilitate novel applications that interact more deeply with the physical world, such as the Internet of Things, autonomous driving, extended reality, smart cities, digital twins [1], and radio maps for environment-aware communication [2]. Thus, to save resources in bandwidth, hardware, processing power, and energy while facilitating efficient operation of such a large number of devices, integrated sensing and communication (ISAC) has been proposed as one of the six key usage scenarios of IMT-2030 (6G) [3]–[6].

73, pp. 4430–4445, 2025

[17] Y. Wan, Z. Hu, A. Liu, R. Du, T. X. Han, and T. Q. Quek, “Ofdm-based multiband sensing for isac: Resolution limit, algorithm design, and open issues,” IEEE Vehicular Technology Magazine, vol. 19, no 2 pp 51–59

[32] R. Bomfin, A. Rasteh, A. Bazzi, H. Guo, H. Lee, M. Mezzavilla, S. Rangan, J Choi, and M Chafii “Multi-Band Channel Sensing in the Upper Mid-Band (FR3)” accepted in IEEE Global Communications Conference (GLOBECOM) 2025 Taipei Taiwan Dec 2025

[41] S. M. Kay, Fundamentals of Statistical Signal Processing, Volume I: Estimation Theory. p. 1

Where the paper says it

Neglecting DMC leads to systematically misleading results: from a system-level perspective, it yields over-optimistic, unrealistic predictions of sensing performance. From a signal processing perspective, it introduces non-negligible estimation bias and unexpectedly large errors, and yields erroneous internal reliability metrics for the estimated parameters.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Background DMC modeling in multi-band ISAC channels allows AI systems to account for frequency-dependent propagation effects like clutter, diffraction, and diffuse scattering by treating them as interference that grows with transmit power. This enables the AI system to produce more meaningful sensing conclusions rather than relying on over-optimistic predictions from SC-only models.

  2. Assessing sensing tradeoffs using Cramer-Rao bound fundamental limits allows the AI system to understand how one sub-band may no longer dominate performance under all conditions, providing a more realistic view of resource allocation across different frequency bands.

  3. A scalable multi-band estimator that resolves angular ambiguities arising from the grating lobes effect allows the AI system to achieve substantial gains in estimation accuracy and reductions in false alarm rate over single-band estimators. This capability is crucial for accurate localization when using signals from communication protocols like FR3.

  4. The proposed algorithm, based on combining single-band estimators with a weighting scheme derived from CRB, enables the AI system to perform a minimal variance linear unbiased estimator by selecting sub-bands where the estimation signal-to-noise ratio is high enough to trust the estimate.

  5. The ESNR based sub-band selection scheme allows the AI system to dynamically drop uncertain sub-bands, preventing combining from harming estimation accuracy in low SNR regions and ensuring that only reliable path estimates are used for localization.

  6. The Bistatic Geometric Gain metric provides a physically meaningful and interpretable proxy for sensing performance by isolating the geometry-dependent interaction between SC, DMC, and noise, enabling resource optimization based on this metric.

Abstract

Multi-band sensing has emerged as a key enabler of integrated sensing and communication (ISAC), one of the six primary usage scenarios defined for IMT-2030 (6G). The introduction of frequency range 3 (FR3, 7-24 GHz), comprising non-contiguous sub-bands across a wide frequency span, further reinforces the importance of multi-band operation. In such scenarios, frequency-dependent propagation effects that are collectively referred to as dense multipath components (DMC), including clutter, diffraction, and diffuse scattering, must be carefully considered. Building on prior literature and our experimental observations, this paper proposes a novel ISAC channel analysis tailored to multi-band sensing, based on a channel model with background DMCs. It also assesses the sensing trade-offs among sub-bands by analyzing Cramér-Rao bound (CRB)-based fundamental limits. Furthermore, a scalable multi-band estimator is proposed that resolves angular ambiguities arising from the grating lobes effect. Simulation results of the multi-band estimator demonstrate substantial gains in estimation accuracy and reductions in false alarm rate over single-band estimators operating on each constituent sub-band within the CRB-achieving regime. In a representative test case, the proposed estimator achieves reductions of 37.41% and 17.04% in the root mean squared error of delay estimation compared to single-band estimators operating at 8.75 GHz and 21.7 GHz, respectively.

Sources

Related papers